liangzulin's recent timeline updates liangzulin's recent timeline updates |
liangzulinV2EX member #327307, joined on 2018-07-06 11:29:40 +08:00 |
liangzulin's recent replies
Apr 28, 2019 Replied to a topic by windflurry › 问与答 › [scala]xgboost4j 中的 train 方法要求接收 dmatrix 参数，怎么将 dataframe 格式转成 dmatrix 啊 |
DataFrame ？你确定说的是 DataFrame 吗？这是 xgboost python 里边的 API 吧。
spark 在 JAVA 里边是 Dataset<>
如果是 pandas DataFrame 的话，转成 DMatrix 是这么转的：
df = pd.read_csv('some_data.csv')
X_train_df = df[['col_x1', 'col_x2', 'col_x3', 'col_x4']]
y_train_df = df[['col_y']]
dtrain = xgb.DMatrix(X_train_df, label=y_train_df)
spark 在 JAVA 里边是 Dataset<>
如果是 pandas DataFrame 的话，转成 DMatrix 是这么转的：
df = pd.read_csv('some_data.csv')
X_train_df = df[['col_x1', 'col_x2', 'col_x3', 'col_x4']]
y_train_df = df[['col_y']]
dtrain = xgb.DMatrix(X_train_df, label=y_train_df)
Select Language